Forced Displacement Cause Area Report
AUTHOR: JOHN HALSTEAD, DPHIL 08/2018
Executive Summary The problem We are currently witnessing one of the worst forced displacement crises in recent memory. By the end of 2016, 65.6m people, almost 1% of the world population, were forcibly displaced as a result of persecution, conflict, violence, or human rights violations. This is predominantly a crisis of low and middle income countries. Since 1991, the vast majority of forced displacement has been caused by the same ten conflicts in the Middle-East, Sub-Saharan Africa, Latin American and Asia. Hosting responsibilities for displaced persons are unevenly shared, with the vast majority residing in the Middle East, Sub-Saharan Africa, and Asia.
Forced displacement is an increasingly urban phenomenon: 50% of forcibly displaced people reside in urban centres and this percentage is expected to increase as low income countries urbanise. 76% of refugees, and 99% of internally displaced persons live outside camps.
Forcibly displaced people include:
• Refugees – displaced people who have crossed a national border
• Internally displaced persons – those who have not crossed a national border
• Asylum seekers
• Stateless persons
• In 2016, there were an estimated 36m internally displaced persons and 17m refugees globally.
Key effects Forced displacement imposes severe costs on those affected. Forcibly displaced people have typically experienced severe trauma in their country of origin. Consequently, they often suffer from serious psychosocial problems: over 30% of people living in conflict-affected regions suffer from post-traumatic stress disorder. Those forcibly displaced often have to leave much of their property behind when fleeing conflict, and have limited socioeconomic rights in host countries. Women face elevated levels of gender-based violence in, during, and after displacement.
2 — Founders Pledge Forced Displacement
Neglectedness The problem of forced displacement is large, pressing, and growing in scale. However, relative to other cause areas, we do not consider it a severely neglected area. We estimate that initiatives related to displacement received upwards of $20bn in 2015. In comparison, several global problems that likely constitute a greater overall burden on humanity (for example Malaria and vitamin A deficiency) receive less funding by an order of magnitude.
Identifying solutions Assessing the impact of policy advocacy is much more time consuming than assessing the effectiveness of direct work. Consequently, in this report, due to time constraints, we restricted our search to charities doing direct work. The evidence on what works to improve the lives of forcibly displaced people is limited, likely because it is difficult to conduct rigorous impact evaluations in fragile and conflict-affected settings. Only a few interventions have been assessed using rigorous methods, and among those that have, many appear to have small effects.
Effective interventions: what works? A review of the existing literature suggests that effective charities doing direct work are likely to do some or all of the following:
• Work in low and middle income countries
• Provide unconditional cash transfers
• Provide layperson-led mental health services
• Provide public health interventions with good evidence from other domains
• Build the evidence-base on interventions targeting the forcibly displaced
• Enable refugees to work legally in high income countries
Charity recommendation: GiveDirectly’s refugee programme Guided by the above considerations, we searched through charities doing direct work on forced displacement, and narrowed down to a shortlist of four charities. GiveDirectly’s refugee programme stood out because it carries out an impactful programme well-supported by evidence, has an outstanding track record and team, and can productively absorb significantly more funds.
3 — Founders Pledge Forced Displacement
GiveDirectly has been delivering unconditional and unrestricted large cash transfers to the world’s poorest people since 2009, and is one of our recommended charities in the area of economic empowerment. In 2017 they started a refugee programme. The first projects focus on Uganda, home to around 1.5m refugees as of 2017, the most of any country in Africa, and the third most in the world. Most of these refugees have fled conflict, persecution and hunger in DRC, South Sudan and Burundi.
We believe that GiveDirectly is one of the most promising charities doing direct work in this area. Cash transfers are one a small number of interventions with some rigorous supporting evidence in the humanitarian context. GiveDirectly has consistently demonstrated the ability to deliver cash transfers in a highly efficient manner – with a pilot study among Ugandan refugees showing that 83% of every dollar donated went to recipients – while having a outstanding commitment to transparency and impact evaluation. We believe that their refugee programme could contribute significant insights to the field as a whole and inform effective strategies in other settings affects by forced displacement.
Direct impact
GiveDirectly is currently seeking funding for a scale-up targeting the entire refugee community in Kiryandongo, Uganda. In order to improve inter-community relations, the project will also target surrounding host community households. We believe that this project could productively absorb an additional $7.5m in the coming year. Each additional $1m would enable them to almost double the annual income of a further ~800 refugee households.
Leverage
Even though there is good reason to believe that cash transfers are more cost-effective than transfers of in-kind goods (such as food and clothes), today it appears that much more humanitarian aid is given in the form of in-kind aid than cash. GiveDirectly’s large trial of cash transfers for refugees has the potential to change the approach taken by the whole humanitarian sector, influencing some portion of the humanitarian aid budget of more than $20bn.
While we are confident in the positive effects of GiveDirectly’s programme, we do think that programmes aiming to improve the health of the global poor are likely to be more impactful, largely due to the relative to neglectedness of global health. People wishing to learn more should see our research page at www.founderspledge.com/research.
4 — Founders Pledge Forced Displacement
Acknowledgements For discussion, advice and expertise, we are grateful to Gary Brittenham, Luz María De Regil, Karen Jacobsen, Susanna Morrison-Metois, Ami Neuberger, Eleanor Ott, Katarina Polonsky, Wietse Tol, Peter Ventevogel, and Helen Verdeli. We are especially grateful to Artemis Seaford for extensive comments on a draft of this report. Any mistakes are our own.
5 — Founders Pledge Forced Displacement
Table of Contents
Executive Summary 2
The problem 2
Key effects 2
Neglectedness 3
Identifying solutions 3
Effective interventions: what works? 3
Charity recommendation: GiveDirectly’s refugee programme 3
Acknowledgements 5
Table of Contents 6
1. Overview of Forced Displacement 8
1.1. Trends in forced displacement 8
1.2. A low-income country crisis 10
1.3. The costs of forced displacement 15
1.4. Is forced displacement a neglected problem? 17
2. Interventions: What Works? 20
2.1. Are there impactful interventions? 20
2.3. Selecting interventions 30
3. Charity Recommendation: GiveDirectly’s Refugee Programme 32
Summary 32
6 — Founders Pledge Forced Displacement
What do they do? 33
Is there evidence the intervention works? 34
Is the intervention cost-effective? 35
What are the other benefits? 39
Is it a strong organisation? 40
Is there room for funding? 40
Risks and reservations 41
Appendix 1. Our process 43
7 — Founders Pledge Forced Displacement
1. Overview of Forced Displacement 1.1. Trends in forced displacement By the end of 2017, 71.4m people, almost 1% of the world population, were forcibly displaced as a result of persecution, conflict, violence, or human rights violations.1 That was an increase of almost 6m people over the previous year.2 The forcibly displaced include the following categories:
• Refugees: Persons crossing an international border fleeing a violent situation seeking international protection.
• Asylum seekers: Persons seeking international protection but whose refugee status has yet to be determined.
• Internally displaced persons: Persons who have fled their homes following violence, but who have remained within their own country.
• Stateless persons: Persons who are not considered as nationals by any State.
Internally Displaced Persons (IDPs) are the most numerous group – numbering 39 million – followed by refugees at 19m.3
1 http://popstats.unhcr.org/en/overview. 2 http://popstats.unhcr.org/en/overview. 3 http://popstats.unhcr.org/en/overview.
8 — Founders Pledge Forced Displacement
Figure 1.
The forcibly displaced population
Source: World Bank, Forcibly Displaced, 2017: p. 16
Figure 1 suggests that forced displacement is now the highest it has been since the Second World War, though historical comparisons should be drawn with caution, as reliable time series data do not exist.4 Nonetheless, the data do suggest that we are currently experiencing one of the most severe forced displacement crises in recent history.5
There has been an especially marked increase in forced displacement since the mid 2000s, driven mainly by the Syrian conflict, but also by other conflicts in the region such as in Iraq and Yemen, as well as in sub-Saharan Africa including Burundi, the Central African Republic, the Democratic Republic of the Congo, South Sudan, and Sudan.6
4 World Bank, “Forcibly Displaced: Toward a Development Approach Supporting Refugees, the Internally Displaced, and Their Hosts,” June 2017, 16. 5 World Bank, 16. 6 UNHCR, “Global Trends,” 5.
9 — Founders Pledge Forced Displacement
1.2. A low-income country crisis Forcibly displaced populations overwhelmingly come from, and reside in, low and middle income countries, rather than high income countries.
Origins of displaced persons
According to the World Bank:
“The large majority of people displaced by conflict do not have the resources or opportunities to flee beyond neighboring areas. They have to remain internally displaced or cross borders in the region.”7
The vast majority of forced displacement has been caused by the same ten conflicts in developing countries since 1991.8 In South Asia and the Middle East, these include prolonged conflicts in Afghanistan and Iraq and the more recent Syrian crisis; in Africa, persistent conflict and instability in Burundi, the Democratic Republic of Congo, Somalia, and Sudan; in Latin America, four decades of internal armed conflict in Colombia; and in Europe and Central Asia, wars in the Caucasus and the former Yugoslavia.9 Figure 2 shows the largest contributors to forced displacement by source country, as of 2016.
7 World Bank, “Forcibly Displaced,” 18. 8 World Bank, 21ff. 9 World Bank, 21ff.
10 — Founders Pledge Forced Displacement
Figure 2.
Forcibly displaced population, by territory of origin
Source: UNHCR, Global Trends, 2016: p. 9.
11 — Founders Pledge Forced Displacement
Hosts of displaced persons Hosting responsibilities for displaced persons are unevenly shared. As Figure 3 shows, the vast majority of displaced persons reside in Sub-Saharan Africa and the Middle East, with a significant portion in Columbia due to the long running conflict involving the FARC militia group.
Figure 3.
Refugees and IDPs, by host country
Source: UNHCR, Population Statistics
As of end-2015, three of Syria’s neighbours (Turkey, Lebanon, and Jordan) hosted 27 percent of all refugees worldwide; two of Afghanistan’s neighbours (Pakistan and the Islamic Republic of Iran) 16 percent; and two of Somalia’s and South Sudan’s neighbours (Ethiopia and Kenya) 7 percent.10 Given their size, some countries have taken an unusually large number of refugees (see Figure 4).
10 World Bank, 23.
12 — Founders Pledge Forced Displacement
Figure 4.
Refugees, as a percentage of host country population
Source: World Bank, Forcibly Displaced, 2017: p. 24.
Contrary to some common perceptions, the number of refugees in EU countries is not only small in relative terms, but also below the peak of the early 1990s.11
An increasingly urban phenomenon
While one might naturally assume that forcibly displaced people predominantly reside in refugee camps, in fact most displaced people reside outside camps. Around 76% of refugees, and 99% of IDPs live outside camps.12 Around 50% of refugees and IDPs live in urban centres. This is predominantly driven by displaced populations in middle income countries (see Figure 5). This proportion is expected to increase as low income countries urbanise.
11 World Bank, 18. 12 World Bank, 26ff.
13 — Founders Pledge Forced Displacement
Figure 5.
Rural/urban location of displaced persons, by region
Source: World Bank, Forcibly Displaced, 2017: p. 28.
The duration of displacement
The longer displacement continues, the greater the burden on displaced persons. Some of the most salient facts about the duration of displacement are as follows:13
• For people who are currently refugees, the mean duration of exile stands at 10.3 years.
• The median duration is 4 years, i.e. the typical person is displaced for 4 years.
• Since 1991, the average duration has fluctuated between about eight years in 1991 and a peak of fifteen years in 2006.
• The number of refugees in protracted situations (five years of exile or more) has been fairly stable since 1991, at 5 to 7 million. For this group, the average duration of exile reached 21.2 years (and the median 19 years), though this is largely influenced by the situation of Afghan refugees.
13 World Bank, 25.
14 — Founders Pledge Forced Displacement
Comparable data is not available for IDPs.14
1.3. The costs of forced displacement Forced displacement imposes numerous costs on those affected. Forcibly displaced people have typically experienced severe trauma in their country of origin. For example, in the Central African Republic, nearly half of the displaced have had a direct experience of violence and more than a fourth have witnessed killings.15 Consequently, the displaced often suffer from serious psychosocial issues.16 Over 30 percent of people living in conflict-affected regions suffer from post-traumatic stress disorder.17
Displaced people often have to leave much of their property behind when fleeing conflict. As the World Bank comments:
“The poverty impact of such losses can be shattering. For example, IDP households in Uganda experienced a 28 to 35 percent decrease in consumption, as well as a significant decrease in the value of their assets compared with non-displaced households. These effects were still felt two years after displacement, and there was no recovery for the bottom quartile households, who appeared to be trapped in poverty. In Colombia IDP household consumption and income fell by 53 and 28 percent respectively, taking the majority of the displaced below the extreme poverty line. In Afghanistan, a study found a 37 percent decrease in ownership levels among Afghan IDPs, and even after five years in displacement, 61 percent of them remained in temporary housing, wracked by insecurity.”18
Forcibly displaced people face major barriers post-displacement. There is currently no exhaustive review of the socioeconomic rights of the forcibly displaced in all large host countries. Nonetheless, in a study of 15 countries by Asylum Access, 45% have a complete legal bar to employment for refugees.19 In the countries where a legal right exists, significant de-facto barriers
14 World Bank, 26. 15 World Bank, 7. 16 Migration Policy Institute, “Building Livelihood Opportunities for Refugee Populations: Lessons from Past Practice,” September 2016, 10. 17 World Bank, “Forcibly Displaced,” 82. 18 World Bank, 81. 19 Asylum Access, “The Global Refugee Work Rights Report,” May 5, 2015, 5, http://asylumaccess.org/global-refugee- work-rights-report/.
15 — Founders Pledge Forced Displacement
to employment, like strict encampment and exorbitant permit fees, undermined refugees’ ability to access sustainable employment.20
The gender aspect of forced displacement
Women face elevated levels of gender-based violence in, during, and after displacement. Rape is prevalent in many of the modern conflicts driving displacement, including Syria, Afghanistan, Somalia, the Democratic Republic of Congo, Sudan, South Sudan, Uganda, Central America, Myanmar, and Nepal.21 In the journey from conflict situations, women and girls risk gender-based violence perpetrated by smugglers, strangers, as well as border patrol guards and detention-centre authorities.22
Moreover, according to the International Labour Organization (ILO), one out of every six irregular female migrants is coerced into sex at destination, and it is plausible that the incidence among the forcibly displaced will be roughly comparable.23 Women may also be abused in their own households. In a recent study, almost two-thirds of displaced Afghan women reported domestic violence during displacement, with nearly a third reporting that it occurred often, very often, or every day, and far more frequently than before displacement.24
Effect on host communities
Displaced persons have mixed effects on host communities. Large influxes of refugees can lead to social tensions and to detrimental economic effects for some sections of society. Opinion polls suggest that there is hostility to refugees in many recipient countries.25 Refugees may also lead to some short-term economic costs as they integrate into informal labour markets.26 For example, in western Tanzania in the mid-1990s, subsistence farmers were unable to compete with refugees in the labour market, and wages paid to casual labourers dropped by up to 50 percent in some
20 Asylum Access, 5. 21 World Bank, “Forcibly Displaced,” 84. 22 World Bank, 84. 23 World Bank, 84. 24 World Bank, 84–85. 25 World Bank, chap. 3. 26 World Bank, 66.
16 — Founders Pledge Forced Displacement
areas.27 Initially accommodating and integrating refugees can be costly, and usage of state benefits is initially high.28
However, refugees also bring some benefits to host countries. Many refugees are highly skilled. For example, it is estimated that nearly half of all Syrian refugees to enter Europe have a university degree.29 In the UK, training a doctor from scratch costs roughly £250,000, whereas certifying a refugee doctor is estimated to cost only £25,000.30 Refugees are also young, which can be useful to recipient countries with ageing populations.31 In some cases, refugees have a positive effect on local economies. For example, the abundance of refugee labour in the Karagwe district of western Tanzania enabled farmers to expand and increase production: between 1993 and 1996 cultivated areas doubled, as did banana and bean harvests.32 In Guinea, the presence of Liberian refugees made it possible to push out rice cultivation to the lower swamp areas.33
1.4. Is forced displacement a neglected problem? When deciding whether and how to work on a problem, it is important to consider not only how important that problem is, but also how neglected it is. For problems that already receive a large amount of resources, the “low hanging fruit” are likely to have been taken, and we may already be experiencing diminishing returns. It may therefore be easier to make progress on more neglected problems.
Forced displacement as a problem appears not to be neglected relative to other important problems, though it is difficult to figure out how neglected forced displacement is. According to the World Bank, “there is no comprehensive picture of the resources provided to support forcibly displaced persons and host communities”.34 No global data are available on the share of humanitarian resources that benefit forcibly displaced persons and their hosts,35 though there is data on humanitarian assistance, much, though not all, of which is likely to be directed towards the
27 World Bank, 67. 28 Owen Barder and Euan Ritchie, “Refugees Are a Boon Not a Burden: Here’s How to Get the Best Results for Everyone,” Center For Global Development, accessed June 5, 2018, https://www.cgdev.org/blog/refugees-boon-not-burden-how- get-best-results. 29 Barder and Ritchie. 30 Barder and Ritchie. 31 Barder and Ritchie. 32 World Bank, “Forcibly Displaced,” 67. 33 World Bank, 67. 34 World Bank, 125. 35 World Bank, 128.
17 — Founders Pledge Forced Displacement
forcibly displaced.36 Humanitarian assistance has grown rapidly over the last 15 years, from $7.2 billion in 2000 to around $27 billion in 2015.37
In-donor refugee costs – costs incurred by host countries on refugees residing in their territories - have risen rapidly since the refugee crisis (see Figure 6). These costs are much higher in high income countries, averaging about $13,000 per refugee per year, whereas low and middle income countries spend on the order of hundreds of dollars per refugee per year.38
Figure 6.
Net Overseas Development Aid on in-donor refugee costs
Source: World Bank, Forcibly Displaced, 2017: p. 128.
We can infer from this that spending on the forcibly displaced comprises the vast majority of the $27bn spent on humanitarian assistance: OECD in-donor refugee costs alone are $12bn, and spending by low and middle income countries on the forcibly displaced is also likely to be in the billions. We should therefore expect that >>$12bn is spent on humanitarian assistance for forcibly displaced people and their hosts.
36 For instance, some humanitarian assistance is spent on things such as the Ebola crisis. 37 World Bank, “Forcibly Displaced,” 126. 38 World Bank, 128.
18 — Founders Pledge Forced Displacement
This suggests that forced displacement is relatively not neglected compared to other important problems. Consider the following examples:
Malaria In 2015, there were an estimated 212m cases of malaria and 429,000 malaria deaths.39 In 2015, malaria was responsible for 55.8m Disability Adjusted Life Years40 (DALYs) (a health metric used to measure the burden of disease – loosely speaking, one DALY averted corresponds to one year of healthy life). However, in 2016, only around $2.6bn of development assistance for health was spent on malaria prevention and treatment41 – 5-10% of the development assistance likely spent on the forcibly displaced.
Vitamin A deficiency Similarly, an estimated 250m schoolchildren alone are vitamin A deficient, and it is likely that in vitamin A deficient areas a substantial proportion of pregnant women is vitamin A deficient.42 An estimated 250,000 to 500,000 vitamin A-deficient children become blind every year, half of them dying within 12 months of losing their sight. 43 In 2015, vitamin A deficiency was responsible for 232m DALYs.44 However, in 2016 nutrition as a whole area received only around $1bn in Overseas Development Assistance. 45
Even though malaria and vitamin A deficiency appear to affect more people than forced displacement and seem comparably severe, they receive only a fraction of the money. Moreover, there are evidence-backed cost-effective interventions in these areas,46 whereas the evidence on interventions on forced displacement is much more scarce. This is some indication that, even though it is important, forced displacement may not be the highest impact problem area to work in.
39 WHO, “Malaria,” accessed June 5, 2018, http://www.who.int/gho/malaria/en/. 40 Nicholas J. Kassebaum et al., “Global, Regional, and National Disability-Adjusted Life-Years (DALYs) for 315 Diseases and Injuries and Healthy Life Expectancy (HALE), 1990–2015: A Systematic Analysis for the Global Burden of Disease Study 2015,” The Lancet 388, no. 10053 (October 8, 2016): 1609, https://doi.org/10.1016/S0140-6736(16)31460-X. 41 Institute for Health Metrics and Evaluation, “Financing Global Health 2017: Funding Universal Health Coverage and the Unfinished HIV/AIDS Agenda,” 2017, 54. 42 WHO, “Micronutrient Deficiencies - Vitamin A,” WHO, accessed September 12, 2014, http://www.who.int/nutrition/topics/vad/en/. 43 WHO. 44 Kassebaum et al., “Global, Regional, and National Disability-Adjusted Life-Years (DALYs) for 315 Diseases and Injuries and Healthy Life Expectancy (HALE), 1990–2015,” 1610. 45 Institute for Health Metrics and Evaluation, “Financing Global Health 2017:,” 54. 46 Donors wishing to learn more should go to www.founderspledge.com/research
19 — Founders Pledge Forced Displacement
2. Interventions: What Works? In this section, we discuss the evidence-base surrounding different interventions that could help the forcibly displaced. This discussion helps to guide the search for impactful charities.
2.1. Are there impactful interventions? There are two possible ways for philanthropists to approach the problem of forced displacement. Firstly, one could tackle the root cause of displacement by trying to reduce conflict, violence and human rights abuses. This would have to be done in an indirect way through research, policy advocacy, education or grassroots activism. Secondly, one could try to respond to the symptoms of the problem by ensuring that displaced people receive adequate aid and are able to live prosperous, secure, happy and healthy lives. This second approach could involve indirect work such as research and policy advocacy, as well as direct work.
In this report, we decided not to investigate research or policy advocacy because we would not have had time to form a reliable judgements on the effectiveness of different organisations given our time constraints.47 To confidently attribute a particular change to a particular policy organisation, one has to make contacts in the field with in-depth knowledge of the influence of particular organisations, and also to understand the political context in great depth, both of which are time consuming. Research and policy organisations also pursue highly heterogeneous projects, making it difficult to infer future success from past success. In contrast, assessing direct interventions is relatively straightforward. Charities carrying out direct interventions tend to implement relatively homogeneous interventions, and it is much easier to attribute a particular change to a specific organisation doing direct work.
Since we have decided to focus on charities doing direct work, and since the vast majority of forcibly displaced people are in low and middle income countries, among charities that do direct work, those that work in low and middle income countries are likely to be more impactful because the scale of the problem there is greater. Moreover, as low and middle income countries are resource-poor, the best opportunities to help refugees may not have been taken, so diminishing returns have not yet set in.
47 We did initially look at some organisations doing policy work, including Overseas Development Institute, Migration Policy Institute, and the Center for Global Development.
20 — Founders Pledge Forced Displacement
The overall state of the evidence
Overall, the evidence in this area is poor. There is little evidence about what works to improve the lives of the forcibly displaced. Only a few interventions have been assessed using rigorous methods, and among those that have, many appear to have small effects. This does not mean that the interventions carried out in the field are ineffective: absence of evidence is not evidence of absence. Indeed, it may be especially difficult to conduct rigorous impact evaluations such as RCTs in fragile and conflict-affected settings. However, our aim here is to analyse the evidence that does exist and to identify high impact evidence-based opportunities.
In our review of the literature, we looked at the following systematic reviews:
• Ott and Montgomery “Interventions to Improve the Economic Self-Sufficiency and Well- Being of Resettled Refugees: A Systematic Review,” Campbell Systematic Reviews 11, no. 4 (2015).
• Crawford et al, Protracted displacement: uncertain paths to self-reliance in exile, Overseas Development Institute, DFID Knowledge for Development, September 2015.
• Jacobsen and Fratzke, Building Livelihood Opportunities for Refugee Populations: Lessons from Past Practice, Migration Policy Institute, 2016.
• McLoughlin, Sustainable livelihoods for refugees in protracted crises,
• Ott, Review of impact evidence for interventions targeting forcibly displaced populations, World Bank, (forthcoming).
The Campbell Review looked at interventions focused on improving the lives of resettled refugees – those that are moved from a host country to a third country, usually a high income one, such as the US, Australia or Canada. The Campbell Review’s inclusion criteria for eligible studies were as follows:
“(a) included a prospective, controlled methodology such as randomised controlled trial design, a quasi-randomised controlled trial design, or a nonrandomised controlled design which provided information on, and adjusted for, baseline comparability; (b) included participants who were refugees who had been served by a refugee resettlement entity and were between the ages of 18 and 64 at the time of the intervention; (c) evaluated an intervention designed to increase the economic self-sufficiency and well-being of resettled
21 — Founders Pledge Forced Displacement
refugees compared to a control or comparison group; and, (d) included at least one primary or secondary outcome (labour force participation rate, employment rate, use of cash assistance, income, job retention, and quality of life).”48
No studies met the inclusion criteria.
The Migration Policy Institute report examined interventions to improve the livelihoods of refugees in both high and low income contexts. In reviewing the state of the evidence, they state:
“Despite the growing interest in and resources devoted to livelihood programs, as yet there is little concrete evidence that current strategies are successfully meeting their goals of fostering self-reliance and durable solutions. In general, there is a lack of independent evaluations, hard data, and external assessments of most livelihoods programs. In fact, two recent review studies found “a near-complete absence” of livelihood evaluations, with the few existing evaluations focused on small-scale NGO interventions and largely qualitative in nature (just 20 percent reflected quantitative research).”49
Later, they note that the few studies that do exist suggest that “livelihood programs have struggled to achieve their stated mission in refugee situations”.50
Similarly, in her review of the evidence on interventions aiming to provide sustainable livelihoods for refugees in protracted crises, McLoughlin comments that “the evidence base is weak both in terms of its size and quality”.51
In a Rapid Evidence Review for the World Bank, Ott (forthcoming) reviewed the literature to find high quality impact evaluations of interventions targeting the forcibly displaced in low, middle and high countries.52 This review paints a more optimistic picture of the evidence-base, with 28 studies meeting the inclusion criteria, as well as 45 additional studies that did not focus on displaced populations but rather on contexts affected by forced displacement. Given the amount of funds
48 Eleanor Ott and Paul Montgomery, “Interventions to Improve the Economic Self-Sufficiency and Well-Being of Resettled Refugees: A Systematic Review,” Campbell Systematic Reviews; Oslo 11, no. 4 (2015): 8. 49 Migration Policy Institute, “Building Livelihoods,” 11–12. See also Overseas Development Institute, Protracted Displacement: Uncertain Paths to Self-Reliance in Exile, 2015; Richard Mallett and Rachel Slater, “Livelihoods, Conflict and Aid Programming: Is the Evidence Base Good Enough?,” Disasters 40, no. 2 (n.d.): 226–45, https://doi.org/10.1111/disa.12142. 50 Migration Policy Institute, “Building Livelihoods,” 12. 51 Claire Mcloughlin, “Sustainable Livelihoods for Refugees in Protracted Crises,” 2017, 2. 52 Eleanor Ott, “Review of Impact Evidence for Interventions Targeting Forcibly Displaced Populations” (World Bank, forthcoming).
22 — Founders Pledge Forced Displacement
flowing into forced displacement, it is surprising that there are only 28 high quality studies on all interventions trying to benefit the forcibly displaced. Ott (forthcoming) notes that “given the limited range of studies that met inclusion criteria, there is a need for more and better-quality rigorous evaluations across a variety of intervention areas”.53
Encouragingly, the evidence base does appear to be growing. Around 90% of the included evaluations were published in 2008 or later, and around half of the evaluations were published in the past 5 years.54 In addition, Ott (forthcoming) identified a number of ongoing and unpublished studies that will address many key gaps in the literature.55
Which interventions work?
The literature that does exist suggests that interventions in the following areas may be effective:
• Cash transfers
• Mental health support
• Giving refugees the right to work in high income countries
There is much less evidence on interventions in the following areas. For each of these areas, Ott (forthcoming) found only unpublished studies, and/or at most one published study reporting on outcomes for forcibly displaced populations.
• Gender-based violence
• Child protection interventions
• Employment and livelihoods
• Education
• Community-driven development interventions
This suggests that we should favour interventions that work on cash transfers or mental health, or try to build the evidence-base on other effective interventions. This relies on the assumption that the only relevant evidence tests interventions that specifically target the forcibly displaced.
53 Ott, 12. 54 Ott, 74. 55 Ott, 74.
23 — Founders Pledge Forced Displacement
However, it might be the case that some interventions that target non-displaced populations are also likely to be effective for forcibly displaced populations. In technical terms, we need to consider the external validity of studies that do not target forcibly displaced populations for forcibly displaced populations. ‘External validity’ refers to the generalisability of findings from one context to another.
The external validity of different findings is likely to differ across interventions. Findings about health interventions, such as treatment for malaria and diarrhoea, are likely to have high external validity for forcibly displaced populations because the mechanism of treatment is largely biological, which we should expect to be fairly similar across contexts. We should expect findings about mental health to have somewhat lower external validity. The mental health challenges faced by displaced populations are likely to be different to those faced by other populations: forcibly displaced populations have suffered trauma at source, have fled their home country, and face major challenges in their host country. Nonetheless, it seems plausible that many of the treatments used for those dealing with trauma in non-displaced populations will sometimes be applicable to displaced populations.
Educating displaced children is also likely to pose substantially different and novel challenges to educating non-displaced children. For example, displaced children are likely to speak a different language and face major psychosocial challenges that could impede learning. Thus, we should expect low external validity in this domain.
It seems plausible that interventions to reduce gender-based violence are likely to have low external validity, as displaced populations are in a relatively unique context with respect to gender dynamics. For example, as discussed above, intimate partner violence increases dramatically in displacement contexts. Finally, interventions targeting livelihoods and basic needs also seem relatively unlikely to generalise well from non-displaced to displaced populations. As discussed above, displaced populations often lack the same socioeconomic rights as the host population and face other forms of prohibitive discrimination.
This suggests that, with the exception of health interventions, evidence from displaced populations is likely to be most relevant when choosing interventions helping forcibly displaced populations. Evidence from other contexts should of course be considered, but concerns about generalisability need to be borne in mind.
24 — Founders Pledge Forced Displacement
The current evidence on cash transfers, mental health, and the right to work in rich countries
In this section we briefly review the evidence on cash transfers and mental health interventions in forced displacement contexts, and on giving refugees the right to live and work in rich countries.
Cash transfers The UNHCR and the World Food Program have used cash-based interventions to replace or supplement food assistance with direct cash transfers and vouchers.56 Between 2000 and 2015, the number of UNHCR country operations utilising cash assistance expanded from 15 to 60 with a total budget of $465m.57 However, while cash transfer programs have become an increasingly important part of social protection programs worldwide, a majority of welfare transfers in both developed and developing countries are still “in-kind” transfers in the form of food or other goods such as clothes.58 In global humanitarian aid in particular, as of 2016, only 10% of aid was in the form of cash.59 As displaced populations increasingly reside in urban areas with better access to markets, the case for cash appears to improve.
There is some evidence regarding the effectiveness of cash transfers for forcibly displaced populations. Ott (forthcoming) found three studies meeting the inclusion criteria that targeted displaced populations.
Table 1 summarises the three studies meeting the Ott (forthcoming) inclusion criteria on unconditional cash transfers in forced displacement contexts.
56 Migration Policy Institute, “Building Livelihoods,” 7. 57 Migration Policy Institute, 7. 58 Jenny C. Aker, “Comparing Cash and Voucher Transfers in a Humanitarian Context: Evidence from the Democratic Republic of Congo,” The World Bank Economic Review 31, no. 1 (February 1, 2017): 44, https://doi.org/10.1093/wber/lhv055. 59 International Rescue Committee, “Seven Steps to Scaling Cash Relief: Driving Outcomes and Efficiency,” March 2018.
25 — Founders Pledge Forced Displacement
Table 1. A summary of the evidence on cash transfers in forcible displacement contexts
Record Intervention Outcomes Context Findings tested
Lehmann & Masterson Cash transfer Impacts of Impacts were measured The average beneficiary household received around $100 per month. (2014) “Emergency (unconditional, cash on for displaced people Around $10 was spent on heating and clothes, and $24 on additional economies: The unrestricted, numerous specifically. food and water. The authors found evidence suggesting that the cash impact of cash targeted to metrics of 1) transfer was sometimes used to cover education costs, with enrolment Country: Lebanon assistance in head of household reaching 39% in the treatment group, compared to 33% in the control Lebanon” household) wellbeing; 2) Context: Syrian refugee group, at the end of the programme. Negative coping strategies were negative households living in also reduced. For example, in the treatment group, around 4%
coping urban settings facing reported sending their children to work, compared to around 10% in Study design: Quasi- strategies; and risks related to winter the control. The study found no evidence of a number of hypothesized experimental, 3) food and negative consequences of cash assistance. For instance, there was no specifically a non-food evidence of beneficiaries spending cash assistance irresponsibly. regression consumption discontinuity design 80% of recipients prefer cash to in-kind aid, though other studies
suggest that such responses may depend on which transfer modality recipients receive.
Aker (2014) Cash transfer To assess the Impacts were measured There were no significant differences in food consumption, wellbeing, “Comparing cash and (unconditional, impact of cash for displaced people total food expenditure, husband making education decisions alone, voucher transfers in a unrestricted, vs. vouchers specifically. deciding whether to share transfer with other households alone, or humanitarian context: targeted to on wellbeing, deciding whether/how to save alone. Since the cost per recipient of Country: Democratic Evidence from the female head of economic the cash program was much lower ($11 vs. $14), the cash programme Republic of the Congo Democratic Republic household) outcomes and was more cost-effective than the voucher programme. of Congo” gender Context: Households in Voucher imbalance. an informal IDP camp in Study design: RCT (restricted to Eastern DRC, during a food or in-kind period of food scarcity goods,
26 — Founders Pledge Forced Displacement
targeted to between September and female head of November household)
Hidrobo et al. (2014a) Unconditional Food security, Impacts measured for While all approaches improved quality and quantity of food consumed, Cash, food, or and dietary both Columbian refugees food vouchers were found to be particularly effective for improving vouchers? Evidence unrestricted diversity, and Ecuadorean non- food quality. In addition, cash transfers were found to be the most from a randomized cash transfer intimate displaced hosts cost-effective modality. Somewhat conflicting survey evidence experiment in targeted to partner suggests that recipients generally preferred cash transfers to other northern Ecuador female head of violence, cost Country: Ecuador modalities, though answers depended strongly on which transfer household effectiveness modality people received. Hidrobo et al. (2014b) Context: Female The effect of cash, Food voucher household members vouchers and food targeted to among Colombian transfers on intimate female head of Refugees and Poor The authors highlight that this effect was observed in an urban area partner violence: household Ecuadorians in an urban with functioning supermarkets, and the impacts may not be the same Evidence from a setting in rural areas. Given that many forcibly displaced persons reside in In-kind food randomized urban areas, though, this is a particularly promising approach. targeted to experiment in Targeting the transfers to female heads of households was found to female head of Northern Ecuador lead to decreases in IPV. household Study design: RCT
Adapted from: Ott, “Review of Impact Evidence for Interventions Targeting Forcibly Displaced Populations”, forthcoming
27 — Founders Pledge Forced Displacement
When deciding whether to support a particular intervention, our aim should be to maximise cost- effectiveness because this ensures we get maximal impact for a given donation. Cost-effectiveness tells us how much it costs in terms of dollars to produce a given benefit for human welfare. The metric used to measure the size of this benefit varies. In global health, benefits are usually measured in terms of DALYs; financial benefits are often measures in terms of the increase in log consumption,60 and so on. For our purposes, it would be useful to answer the following question: are cash transfers more cost-effective than in-kind transfers in forced displacement contexts or in similar settings? While there are a number of papers comparing cash and in-kind aid in terms of other factors, such as “cost to deliver a given bundle of goods” or “operational cost per person reached”, surprisingly few academic research papers have tried to answer this question.
In the forced displacement context, only Hidrobo et al (2014) (described in Table 1 above) has tried to compare the cost-effectiveness of cash compared to in-kind transfers. They found that the cost to produce a given 15% improvement in various food consumption metrics was 2-4x lower for cash compared to in-kind food transfers.61 Although this is informative, this is a restricted welfare metric, as there are other contributors to human welfare than food consumption. Since cash gives people the option of buying a range of goods aside from food, such as clothing, shelter, or medicine, we should expect the Hidrobo et al metric to understate the cost-effectiveness of cash relative to food transfers.
The evidence from humanitarian contexts not affected by displacement is also sparse. It is usually cheaper to get a given cash transfer to a recipient than an in-kind transfer because the operational costs and administrative goods of handling and transporting goods such as food and clothing tend to be higher than delivering cash.62 However, the overall cost of delivering a given amount of food depends on the economic context, and in many cases in-kind aid saves money by buying the goods wholesale rather than in local markets.63 Studies that do calculate cost-effectiveness tend to use restricted welfare metrics relating only to diet, as in Hidrobo et al (2014).64 These produce
60 This measures the welfare benefit of increasing consumption by relativising to the person’s initial income. The idea behind this is that a given amount of additional money is worth more to a poor person than a rich person. 61 Melissa Hidrobo et al., “Cash, Food, or Vouchers? Evidence from a Randomized Experiment in Northern Ecuador,” Journal of Development Economics 107 (March 1, 2014): 152–53, https://doi.org/10.1016/j.jdeveco.2013.11.009. 62 For overviews see Hannah Tappis and Shannon Doocy, “The Effectiveness and Value for Money of Cash-Based Humanitarian Assistance: A Systematic Review,” Journal of Development Effectiveness 10, no. 1 (January 2, 2018): 135–36, https://doi.org/10.1080/19439342.2017.1363804; Courtenay Cabot Venton, Sarah Bailey, and Sophie Pongracz, “Value for Money of Cash Transfers in Emergencies,” Report Prepared for DFID. London: DFID, 2015, 11–12; Sarah Bailey and Paul Harvey, “State of Evidence on Humanitarian Cash Transfers,” Overseas Development Institute Background Note, 2015. 63 Cabot Venton, Bailey, and Pongracz, “Value for Money of Cash Transfers in Emergencies,” 9. 64 Bailey and Harvey, “State of Evidence on Humanitarian Cash Transfers,” 3–4.
28 — Founders Pledge Forced Displacement
mixed results, with several studies showing that cash is more cost-effective but with some showing food to be better.65
The restricted welfare metrics used probably understate the benefits of cash relative to in-kind aid for the reasons outlined above. Since cash gives people the option of buying goods other than food, metrics that focus only on food miss out on the benefits from non-food purchases. Indeed, there is evidence suggesting that recipients of in-kind aid sometimes resell their goods on the local market in order to purchase other goods. For example, 70% of Syrian refugees in Iraq have reportedly resold their in-kind aid;66 while recipients of food aid in Ethiopia resold 30-50% of their wheat entitlement.67
Overall, our view is that cash transfers are likely to be more cost-effective than in-kind transfers provided beneficiaries have access to reasonably well-functioning local markets. Limited existing evidence using restricted welfare metrics suggests that, in certain conditions, cash is generally more cost-effective than in-kind aid. Since these welfare metrics understate the benefits of cash relative to in-kind aid, in the right conditions cash is more cost-effective than in-kind transfers.
Mental health There is evidence supporting the notion that mental health support can be successfully adapted to forced displacement contexts. Interventions targeting adults and children have been shown to reduce mental illnesses, such as depression, anxiety, and PTSD. Ott (forthcoming) found 4 evaluations of interventions carried out by community lay counselors targeting a forcibly displaced population, which found significant decreases in baseline depression symptoms, with some evaluations also reporting decreases in PTSD and anxiety symptoms.68 Interventions carried out by trained mental health professionals have also been shown to be effective in treating depression and PTSD.
In addition, Ott (forthcoming) state that their findings “suggest that there are therapeutic approaches that can be successfully adapted for forcibly displaced children and adolescents”. A
65 Tappis and Doocy, “The Effectiveness and Value for Money of Cash-Based Humanitarian Assistance,” 135–36; Cabot Venton, Bailey, and Pongracz, “Value for Money of Cash Transfers in Emergencies,” 9ff; Ugo Gentilini, “Revisiting the ‘Cash versus Food’ Debate: New Evidence for an Old Puzzle?,” The World Bank Research Observer 31, no. 1 (February 1, 2016): 156–57, https://doi.org/10.1093/wbro/lkv012. 66 Overseas Development Institute and Center for Global Development, “Doing Cash Differently: How Cash Transfers Can Transform Humanitarian Aid. Report of the High Level Panel on Humanitarian Cash Transfers,” September 2015, 18. 67 Cabot Venton, Bailey, and Pongracz, “Value for Money of Cash Transfers in Emergencies,” 21. 68 Ott, “Review of Impact Evidence for Interventions Targeting Forcibly Displaced Populations,” 6.
29 — Founders Pledge Forced Displacement
recent systematic review focused on conflict-affected children in general highlights that most of the effective practices appear to use a combination of “access promotion, psychoeducation for children and parents, insight building, rapport building techniques, cognitive strategies, use of narratives, exposure techniques, and relapse prevention”.69
In the research for our mental health report, we found that interventions delivered by skilled mental health practitioners were unlikely to be cost-effective in low and middle income contexts because such professionals are scarce in those contexts.70 We therefore believe that approaches that make use of “task shifting” – the use of lay community members to lead mental health interventions – are more likely to be cost-effective.
The right to work in rich countries Although the vast majority of refugees reside in low and middle-income countries, one way to potentially have a very large impact on a lower number of refugees is providing them with the right to live and work in rich countries. The ‘place premium’ refers to the gain in income that can be achieved by moving from a poor country to a rich one.71 The evidence suggests that the wages of a poor country worker would increase enormously if they were able to move to a rich country. Controlling for other confounders and selection effects, the place premium for a Bolivian worker would be an increase in wages by a factor of 2.7.72 For a Nigerian, the place premium of a move to the US is 8.4x.73 Thus, the gains to giving refugees from poor countries the right to live and work in rich countries are potentially very large, and might be large enough to counterbalance the relatively limited reach of interventions focusing on the much larger population of displaced people in low and middle-income countries.
2.3. Selecting interventions The foregoing discussion suggests that, to find impactful charities doing direct work, we should search for charities that do some combination of the following:
1. Work in low and middle-income countries.
2. Build the evidence-base on interventions targeting the forcibly displaced.
69 Ott, 6. 70 See our cause report on mental health at www.founderspledge.com/research 71 Michael A. Clemens, Claudio E. Montenegro, and Lant Pritchett, The Place Premium: Wage Differences for Identical Workers across the US Border (The World Bank, 2008). 72 Clemens, Montenegro, and Pritchett. 73 Clemens, Montenegro, and Pritchett.
30 — Founders Pledge Forced Displacement
3. Provide unconditional cash transfers or layperson-led mental health services to forcibly displaced populations or deliver public health interventions with good evidence from other domains.
4. Enable refugees to legally work in high income countries.
These criteria guided our search for cost-effective charities targeting the forcibly displaced. We discuss our process for selecting charities in more depth in Appendix 1. Our process.
It is important to note that since we have excluded policy interventions, our restricted aim is to recommend highly cost-effective charities doing direct work to benefit forcibly displaced populations. We plan to consider organisations doing policy work when we revisit the report in future.
31 — Founders Pledge Forced Displacement
3. Charity Recommendation: GiveDirectly’s Refugee Programme After searching through charities working in the forced displacement space, we recommend GiveDirectly’s refugee programme. This conclusion is based in large part on research by our research partner GiveWell. GiveDirectly is a US-based charity that transfers cash to households in low and middle income countries via mobile phone-linked payment services. In 2018, following a successful pilot in 2017, it started a programme providing large lump sum cash transfers to refugees in Uganda.
Summary
What do they do? Provide unconditional cash transfers to refugees in Uganda and Rwanda.
Is there evidence the intervention works? There is good evidence from forced displacement and non-displacement contexts that unconditional cash transfers provide benefits to recipients.
Is the intervention cost-effective? The intervention is relatively low-cost to deliver, and provides potentially substantial benefits to recipients in extreme poverty.
What are the wider benefits? GiveDirectly is carrying out a randomised control trial of its refugee programme, which should help to expand the evidence-based and inform the practice of development actors. The value of this information could be quite substantial.
Is it a strong organisation? GiveDirectly is an outstanding organisation. It is unusually committed to evidence, impact and transparency. Around 87% of money donated to GiveDirectly goes to the extreme poor.
Is there room for more funding? GiveDirectly’s scaled-up refugee project in Uganda could productively absorb a further $7.5m in funding. Each additional $1m would enable GiveDirectly to reach a further ~800 more refugee households.
32 — Founders Pledge Forced Displacement
What do they do? GiveDirectly transfers cash to poor households in developing countries primarily via mobile phone- linked payment services. It has operated since 2009 and is currently active in Kenya, Uganda, and
Rwanda. To date, GiveDirectly has primarily provided large, lump sum transfers. It recently started a basic income guarantee program, in which recipients will receive long-term (over two or twelve years in the initial study), ongoing cash transfers sufficient for basic needs. It has delivered cash to over 100,000 households in Africa.
GiveDirectly is now running a programme transferring cash to refugees. In 2017, GiveDirectly started a pilot programme transferring cash to refugees in Uganda.74 The pilot programme targeted around 4,400 refugees in protracted exile – i.e. those who fled their homes 5 or more years ago – in Kyaka, Uganda, receiving lump sum cash transfers of around $660 per household.75 The transfers were unconditional and unrestricted, meaning that recipients can spend the money on what they like, and the cash is not conditional on certain actions. The transfers also targeted host communities, with the aim both of providing economic support and strengthening social bonds between refugees and the host populations.76
The pilot study showed that it is operationally feasible and efficient to deliver large, unrestricted cash transfers in refugee settings. 83% of every dollar donated went to recipients in Kyaka.77 Recipients safely and securely received the transfers. For instance, cases of theft were low: in total, the combined loss of transfers to theft and other adverse events was just 0.15% of the total transfer size.78 Opinion poll respondents stated that they strongly preferred cash to in-kind aid, though as discussed in section 2.2, this may be an artefact of the fact that they received cash rather than in- kind aid.79 Early evidence tentatively suggests that there were positive improvements in recipients’ outcomes.80 Evidence from focus groups also weakly suggests that the intervention may have improved relations between host communities and refugees.81
74 GiveDirectly, “Announcing Cash for Refugees,” March 2018, https://givedirectly.org/blog- post?id=1487208189230521130. 75 GiveDirectly, “Review of Uganda Refugee Project Pilot,” June 2018. 76 GiveDirectly, “Announcing Cash for Refugees.” 77 GiveDirectly, “Review of Uganda Refugee Project Pilot.” 78 GiveDirectly. 79 GiveDirectly. 80 GiveDirectly. 81 GiveDirectly.
33 — Founders Pledge Forced Displacement
GiveDirectly plans to carry out a further pilot in Rwanda as well as a high quality RCT or regression analysis of a scaled up project in Uganda. The scaled-up programmes will target 9,000 households in protracted exile – the entirety of a refugee community in Kiryandongo, Uganda –with lump sum cash transfers of $1,000 per household.82 Those in protracted exile are identified as those who are no longer entitled to receive food aid or cash because they had resided in the settlement for more than 3 years.83 The project will also target 3,800 surrounding host community households in order to reduce the risk of tensions.84 If the evaluation shows the scale-up to be a success, then GiveDirectly aims to scale-up further, and also to influence other actors in the space, such as UNHCR.85
GiveDirectly has delivered cash to Uganda since 2013.86 Uganda is now home to the largest number of refugees in Africa, with around 1.35m as of 2017,87 primarily due to conflict and hunger in Burundi, South Sudan and the Democratic Republic of Congo. This is the third highest number of refugees in the world, after Turkey and Pakistan. Uganda has highly progressive refugee policies.88
Is there evidence the intervention works? Unconditional cash transfers have been shown in both forced displacement and non-displacement contexts to provide benefits to recipients.
The evidence from non-displacement contexts is discussed in some depth in the report on cash transfers on our research page authored by our research partner GiveWell.89 Studies generally show substantial increases in short-term consumption, especially food, and little evidence of negative consumption such as tobacco or excessive alcohol consumption.
The evidence most relevant to GiveDirectly comes from an RCT of a GiveDirectly campaign in rural Kenya, which showed that lump sum unconditional cash transfers led to large increases in recipients' consumption, assets, business investment, and revenue.90 There is also some evidence
82 Joe Huston, written correspondence, 1st August 2018. 83 Joe Huston, written correspondence, 10th May 2018. 84 Joe Huston, written correspondence, 12th July 2018. 85 Joe Huston, written correspondence, 10th May 2018. 86 Joe Huston, written correspondence, 10th May 2018. 87 http://popstats.unhcr.org/en/overview 88 World Bank, “An Assessment of Uganda’s Progressive Approach to Refugee Management,” August 2016, http://www.worldbank.org/en/topic/fragilityconflictviolence/brief/ugandas-progressive-approach-refugee- management. 89 See www.founderspledge.com/research 90 Johannes Haushofer and Jeremy Shapiro, “Household Response to Income Changes: Evidence from an Unconditional Cash Transfer Program in Kenya,” Massachusetts Institute of Technology, 2013.
34 — Founders Pledge Forced Displacement
that, in non-displacement contexts, recipients are able to invest cash transfers at high rates of return (e.g., ~20% per year), leading to long-term increases in consumption, though this has recently been questioned (see Risks and reservations).
We discussed the evidence for the forced displacement context in section 2.2, which showed that cash transfers are one of the few interventions with reasonable evidence behind them in a forced displacement context.
Is the intervention cost-effective? GiveWell models the effect of GiveDirectly’s programme by calculating the effect of cash transfers on short-term consumption and from investment returns in the future.91 From this, it calculates the benefits of unconditional cash transfers in terms of a welfare metric: “cost per outcome as good as averting the death of a child under 5”. According to the value and empirical inputs of the median GiveWell staff member, GiveDirectly produces an outcome as good as averting the death of an under 5 for $11,273.92
The opportunities and constraints faced by refugees are likely to differ in some important respects from those faced by non-displaced people served by GiveDirectly. Thus, there are likely to be a number of relevant differences between refugees and GiveDirectly’s usual recipients, which could make this estimate inapplicable. We will now compare the most important features which could affect the cost-effectiveness of the refugee programme as opposed to the “standard programme” not targeting refugee populations.
Size of transfer
The size of the transfer for the scaled-up Uganda project is likely to be roughly the same as the transfer for GiveDirectly’s standard programme. GiveDirectly currently supply grants of around $1,000 per household in their standard programme.93 For the Uganda refugee scale-up, they tentatively plan, on the basis of the pilot, to provide grants of around $1,000 per household.94
91 https://www.givewell.org/how-we-work/our-criteria/cost-effectiveness/cost-effectiveness-models. ‘Cash’ sheet. 92 https://www.givewell.org/how-we-work/our-criteria/cost-effectiveness/cost-effectiveness-models. ‘Results’ sheet, cell B17. As of July 2018. 93 https://www.givewell.org/how-we-work/our-criteria/cost-effectiveness/cost-effectiveness-models. ‘Cash’ sheet. 94 Joe Huston, written correspondence, 10th May 2018.
35 — Founders Pledge Forced Displacement
Thus, the size of the transfer is unlikely to make a large difference to the relative effect of the refugee programme.
Baseline consumption level of recipients
For the standard programme, GiveDirectly’s average recipient has a baseline annual consumption of around $285.92.95 We would guess that refugees in Uganda have roughly comparable consumption to this, though data on this is difficult to come by. A report by the World Bank on Uganda’s refugee situation suggests that Ugandan nationals generally have higher wages than refugees, with refugee wages between ~40% and ~90% of nationals’ annual wages depending on the geographical area.96 However, the income of refugees is likely to be fairly close to that of the households targeted by GiveDirectly’s standard programme, who are poorer than the national average.
The baseline level of consumption matters because it is widely held that money has diminishing marginal benefits for welfare: the poorer you are, the more money improves your welfare. GiveWell accounts for this by modelling the benefits of increased consumption as a function of the natural logarithm of consumption: within a certain range, doubling consumption from any level increases welfare by the same amount. For instance, doubling consumption from $200 to $400 adds as much welfare as doubling income from $1,000 to $2,000. This captures the idea that money has diminishing marginal utility: additional money is more valuable if you are extremely poor than if you are rich. We think it unlikely that the difference in recipient income between the standard programme and the refugee programme would have a major effect on the cost-effectiveness analysis of cash transfers to refugees.
Consumption opportunities
The benefits of cash transfers stem from opportunities to consume and opportunities to invest. We think that consumption opportunities are likely to be broadly similar across recipients of the standard programme and the refugee programme. The World Bank report on Uganda comments:
95 https://www.givewell.org/how-we-work/our-criteria/cost-effectiveness/cost-effectiveness-models. ‘Cash’ sheet. 96 World Bank, “An Assessment of Uganda’s Progressive Approach to Refugee Management,” 65.
36 — Founders Pledge Forced Displacement
“Uganda has progressive legal and policy frameworks that entitle refugees to the right to work, freedom of movement, and access to Ugandan social services.”97
Because refugees have the right to free movement, it seems likely that they will have access to similar markets to those available to standard programme recipients. Indeed, health, education and water services are now integrated across host and refugee populations.98 In light of the widespread opportunities to consume, we would expect that the short-term consumption effects of the transfers will be broadly similar. To model the overall welfare effects of cash transfers, one has to make assumptions about values – the contribution of money to welfare – and about empirical facts, such as how much of the cash transfer is invested. According to the value and empirical assumptions of the median GiveWell staff member, the short-term consumption benefits comprise 55% of the welfare benefit of the programme.99
Investment opportunities
Part of the modelled benefit of GiveDirectly’s standard programme derives from the returns from investments enjoyed by recipients. For example, following one cash transfer, more than 75% of villagers used their transfer to replace their thatched roof with an iron roof.100 Thatched roofs have to be replaced every 3-4 months at significant cost, so the iron roof has potentially large returns, on the order of around 20% p.a.
Refugees enjoy fewer opportunities to invest than standard programme recipients. Refugees enjoy fairly extensive ownership rights in Uganda: they have the right to own and dispose of movable property and the right to lease and sublease immoveable property, such as land. Refugees residing in designated settlements are given reasonable access to land for the purpose of cultivation or pasturing.101 However, the World Bank notes:
“Refugees do not, however, have the right to sell or lease the land allocated to them “strictly for their individual or family utilization.” Refugees residing outside the designated areas may legally acquire leasehold titles, but not freehold, just like other resident aliens, and they may freely dispose of their occupancy interest or sublease on commercial terms.”102
97 World Bank, 27. 98 World Bank, 30. 99 https://www.givewell.org/how-we-work/our-criteria/cost-effectiveness/cost-effectiveness-models. ‘Cash’ sheet. 100 See the report on GiveDirectly at www.founderspledge.com/research 101 World Bank, “An Assessment of Uganda’s Progressive Approach to Refugee Management,” 12. 102 World Bank, 12.
37 — Founders Pledge Forced Displacement
Refugees have relatively good access to employment opportunities. They have the right to: (1) engage in agriculture, industry, and business, whether as workers or proprietors; (2) practice their profession, provided they are properly qualified with recognized certificates; and (3) access formal and informal employment opportunities wherever available in the country and without the need to first obtain work permits.103 Refugees have the right to access employment on a par with the most favoured alien citizen to Uganda.104
These extensive socioeconomic rights have led refugees to be involved in a range of economic activities.
“Enterprises that are mainly run by Congolese, Ethiopian, and Rwandan refugees include small-scale trading of accessories, selling of fabric, retail trading, brokerage services with countries of origin, tailoring, and operating very small restaurants and bars... Refugees throughout Uganda are engaged in similar small businesses, such as tailoring, local brewing, operating restaurants, food vending, selling charcoal, domestic work, retail and whole trade, mobile money businesses, construction, transport, and boda-boda (motorcycle taxi).”105
In spite of this, in surveys refugees cite a number of barriers to successfully attaining employment. ~5-30% of respondents report difficulty finding a job due to unfamiliarity with the language, legal issues, inadequate interviewing skills, discrimination, and a lack of relevant documents.106
GiveWell use returns to iron roofs as a proxy for the returns to investment from cash transfers.107 According to the World Bank study, in Kampala, most refugees live in rental accommodation, but outside Kampala most live in temporary houses made of mud walls and floors with polythene roofs.108 It is unclear whether these houses could be upgraded with iron roofs, or what the returns to this would be.
Overall, the evidence suggests the returns to investment for refugees are likely to be lower than those enjoyed by recipients of the standard programme. Our best, though highly uncertain, guess
103 World Bank, 13. 104 World Bank, 13. 105 World Bank, 35. 106 World Bank, 37. 107 https://www.givewell.org/how-we-work/our-criteria/cost-effectiveness/cost-effectiveness-models “user inputs” sheets. 108 World Bank, “An Assessment of Uganda’s Progressive Approach to Refugee Management,” 60–61.
38 — Founders Pledge Forced Displacement
is that the returns on investment would be around 50% lower than those enjoyed by Ugandan nationals.
An overall assessment
We have seen that the income level of recipients, the size of the cash transfer, and the short-term consumption benefits are unlikely to make a large difference to the cost-effectiveness of the refugee programme. However, the investment benefits may be lower. According to the estimates of the median GiveWell staff member, recipients invest 39% of their cash transfer,109 and returns to investment account for 45% of the benefits of the standard programme.110 If the refugees invest the same proportion of their transfer but receive 50% the return on investment enjoyed by standard programme recipients, then the refugee programme would be (55%+(45%*50%)) = 77% as effective the standard programme. However, since the returns to investment are likely to be lower for refugees, we should expect them also to consume more and invest less with their cash transfer. It is unclear what the magnitude of this effect would be but it seems reasonable to estimate that the refugee programme would be more than 77% as effective as the standard programme.111
If it is 80% as effective, then, using the value and empirical inputs of the median GiveWell staff member, GiveDirectly would produce benefits as good as saving the life of a child under 5 for $14,091. This is highly cost-effective in comparison to many other direct poverty and health interventions.
What are the other benefits? GiveDirectly’s refugee programme could produce potentially highly valuable information about the efficacy of cash transfers in refugee contexts. As discussed in section 2, there are few studies of cash transfers in displacement contexts or in humanitarian contexts more generally. GiveDirectly’s scale-up in Uganda would be the first rigorous trial testing the effectiveness of large lump sum cash transfers: other trials in forced displacement contexts have examined smaller transfers, or transfers delivered on a recurring basis.112 Moreover, there is little good evidence about whether
109 https://www.givewell.org/how-we-work/our-criteria/cost-effectiveness/cost-effectiveness-models. ‘Cash’ sheet, row 5. 110 https://www.givewell.org/how-we-work/our-criteria/cost-effectiveness/cost-effectiveness-models. ‘Cash’ sheet. There is some reason to think this estimate may be too high. See Risks and reservations below. 111 GiveDirectly believes there is a reasonable case to be made that the direct impact could be higher for refugees because they are more credit and capital constrained. Joe Huston, personal correspondence, 14th June 2018. 112 Christian Lehman and Daniel Masterson, “Emergency Economies: The Impact of Cash Assistance in Lebanon,” August 2014; Aker, “Comparing Cash and Voucher Transfers in a Humanitarian Context”; Hidrobo et al., “Cash, Food, or Vouchers?”
39 — Founders Pledge Forced Displacement
any interventions are effective in forced displacement contexts. In light of this, an additional highly powered trial could provide valuable information to non-profits working in the space, insofar as they are motivated by evidence. Since >>$12bn is moved into the forced displacement space every year, a shift in approach brought about by rigorous evidence could have a large positive impact.
GiveDirectly have told us that UNHCR has shown significant interest in the outcome of the trial.113 UNHCR more than doubled its use of cash transfers to $688m in 2016 from $325m in 2015,114 and appears committed to increasing this further in the future. However, this is still only a relatively small fraction of the UNHCR’s overall budget of $7.51bn.115 Thus, there seems to be significant scope for the expansion of this evidence-backed intervention at UNHCR. The GiveDirectly trial could also increase the use of cash among other actors, which remains fairly low in the space, as discussed in section 2.2.
Is it a strong organisation? GiveDirectly is an exceptionally strong and efficient organisation.116 It is committed to self- evaluation, has a strong track record, and has outstanding transparency. GiveDirectly has invested heavily in self-evaluation since its inception, participating in pre-registered high quality RCTs of its programmes. It has repeatedly updated its approach in light of new evidence. It has successfully accomplished its goal of transferring cash to extremely poor households. According to GiveWell, GiveDirectly is one of the most transparent organisations they have encountered.
Is there room for funding? We believe that GiveDirectly’s refugee programme could absorb significantly more funds from summer 2018 to summer 2019, primarily for the Uganda scale-up and evaluation. The minimum budget to enrol the first 50% of recipients for the Uganda scale-up and evaluation is $9m.117 To deliver the cash transfers to the remaining 50%, a further $7.5m would be required, bringing the total budget to $16.5m.118 The budget is divided as follows:119
113 Joe Huston, written correspondence, 10th May 2018. 114 UNHCR, “Global Report 2016,” 2016, 11, http://reporting.unhcr.org/sites/default/files/gr2016/pdf/Book_GR_2016_ENGLISH_complete.pdf. 115 UNHCR, 3. 116 See www.founderspledge.com/research 117 Joe Huston, written correspondence, 1st August 2018. 118 Joe Huston, written correspondence, 1st August 2018. 119 Joe Huston, written correspondence, 1st August 2018.
40 — Founders Pledge Forced Displacement
• Research costs = $1.5m
• Cash transfer delivery costs = $2.2m
• Cash transfers = $12.8m
They currently have banked $1.2m, and given current conversations and match funding, they have a strong pipeline of a further $7.8m-$8.8m,120 meaning that the project has now reached minimum viability. Money beyond this, up to $16.5m will improve the statistical power of the study, and will, of course, increase the direct impact of the project by providing additional cash to refugees. Each additional $1m would enable GiveDirectly to reach a further ~800 more refugee households.
Risks and reservations There are some uncertainties about the cost-effectiveness of GiveDirectly’s refugee programme. Firstly, one natural worry is that the large cash transfers would be a pull factor encouraging more refugees to go to Uganda in the hope of receiving cash transfers. In South Sudan, GDP per capita is $850 as of 2015,121 and in the DR Congo it is $2,350. A lump sum cash transfer of $1,100 per household would provide a substantial fraction of annual income for a four person refugee household. The Lehman and Masterson study for IRC suggests that the cash transfers were not a significant pull factor for Syrian refugees to enter Lebanon, but there is little other evidence on this issue.122 We are at present overall uncertain about whether large cash transfer would be a pull factor.
In addition, some doubts have recently been raised about the cost-effectiveness of GiveDirectly’s standard programme.123 Firstly, there is the risk of negative spillovers. A three year follow up study by Haushofer and Shapiro has shown that cash transfers may have substantial negative effects on non-recipients who live near recipients (“negative spillovers”), in terms of reduced expenditure and food security.124 Under one interpretation, these effects are large enough to outweigh the positive
120 Joe Huston, written correspondence, 10th May 2018. 121 https://data.worldbank.org/country/south-sudan 122 Lehman and Masterson, “Emergency Economies,” 36. 123 For the initial critique see Berk Ozler, “Dear Governments: Want to Help the Poor and Transform Your Economy? Hold On, Recalculating…,” Text, World Bank Blogs, March 26, 2018, http://blogs.worldbank.org/impactevaluations/dear- governments-want-help-poor-and-transform-your-economy-hold-recalculating; Berk Ozler, “GiveDirectly Three-Year Impacts, Explained,” Text, World Bank Blogs, March 30, 2018, http://blogs.worldbank.org/impactevaluations/givedirectly-three-year-impacts-explained.For an overview see GiveWell, “New Research on Cash Transfers,” The GiveWell Blog, May 4, 2018, https://blog.givewell.org/2018/05/04/new-research- on-cash-transfers/. 124 Ozler, “Dear Governments.”
41 — Founders Pledge Forced Displacement
effects of cash transfers. GiveWell’s best guess is that they will revise their estimate of the cost- effectiveness of cash transfers to some extent, but will likely continue to recommend GiveDirectly.125 They plan to investigate the issue further in the middle of 2018. GiveDirectly is currently carrying out an RCT that is designed to measure the spillover effects of GiveDirectly’s programme, and we will have more information on that in the next few months.
The risk of negative spillovers is likely to be lower in the Kiryandongo scale-up because the current plan is to enrol the entire refugee settlement as well as neighbouring host communities.126 Since no nearby communities are excluded from benefits, the risk of negative spillovers is likely to be reduced. In order to provide a control population, half of the recipients will receive payments at the start of the programme and half two years later.127 There is some risk that those randomised to receive the cash transfers later on, or non-recipients in other host communities could experience negative spillovers. This will be revealed in the evaluation of the scale up, and more will be revealed by GiveDirectly’s study of non-refugee populations mentioned above. Finally, as discussed above, evidence from the Kyaka pilot weakly suggests that the approach of targeting both the refugee and host population reduced tensions between the two populations.
The final potential risk with GiveDirectly is the concern that the effects of cash transfer programmes fade out over time.128 Several new studies seem to find that cash may have little effect on recipients’ standard of living beyond the first year after receiving a transfer. GiveWell expects to revise its cost-effectiveness estimate of GiveDirectly down somewhat as a result of this. Because most of the modelled benefits of cash derive from short-term increases in consumption, if the long- term effects are zero, this would lead us to revise the cost-effectiveness estimate of cash transfers down by about 45%. However, for the reasons outlined above, refugees are likely to invest less of their transfer anyway, so this would not make a large difference to our rough cost-effectiveness estimate of transfers to refugees outlined above. GiveWell plans to review this new research in more detail later in the year. This research would inform our charity recommendations in this space going forward.
125 GiveWell, “New Research on Cash Transfers.” 126 Joe Huston, email correspondence, 12th July 2018. 127 Joe Huston, email correspondence, 12th July 2018. 128 Ozler, “GiveDirectly Three-Year Impacts, Explained”; Ozler, “Dear Governments.”
42 — Founders Pledge Forced Displacement
Appendix 1. Our process Our cause reports use a top-down approach to charity selection. We begin at a high level by gaining an understanding of the area. Next we survey the academic literature on interventions that may be effective in the space. We then search for charities implementing these interventions, and recommend them according to various criteria including internal monitoring, track record and room for more funding.
Our review of forced displacement and of the academic literature on effective interventions suggested that, among charities implementing direct interventions, those that do the following are likely to be most cost-effective:
1. Work in low and middle income countries.
2. Build the evidence-base on interventions targeting the forcibly displaced.
3. Provide unconditional cash transfers or layperson-led mental health services to forcibly displaced populations, or deliver public health interventions with good evidence from other domains.
4. Enable refugees to legally work in high income countries.
We built a longlist of charities that fit these criteria by: using past knowledge of charities working in the space; asking philanthropists and area experts for recommendations; and carrying out a general internet search. A shortlist was generated by fit with the key themes above, availability of impact evaluations, and previous experience with charities.129 Our shortlist included:
• GiveDirectly’s Refugee Programme
• International Rescue Committee
• International Refugee Assistance Project
• Sanku Project Healthy Children
We will now discuss why we did not recommend the International Rescue Committee, the International Refugee Assistance Project, or Sanku Project Healthy Children.
129 Before deciding to deprioritise organisations chiefly working on policy, we also briefly looked at Overseas Development Institute, Migration Policy Institute, and the Center for Global Development.
43 — Founders Pledge Forced Displacement
International Rescue Committee
One of the charities we looked at in some depth was the International Rescue Committee (IRC). IRC is a large international NGO that provides direct assistance in humanitarian crises, builds the evidence-base on effective interventions in that space, and advocates for policies to benefit refugees in high income countries and in low and middle-income countries. As discussed, we did not assess their policy work in depth for this report, though we hope to do so for the updated refugee report next year.
We are positive about IRC chiefly because they are responsive to evidence and have done more than any other NGO to build the evidence-base for what works in humanitarian crises. One of their key stated aims is to test the impact of their programmes and to scale the most effective ones.130 Their responsiveness to evidence is illustrated by the fact that, in light of the evidence on the cost- effectiveness of cash transfers, they increased the proportion of humanitarian aid spent on cash transfers from 6% in 2015 to 17.7% in 2017, well above the average of 10% in the humanitarian space.131 They aim to reach 25% by 2020.132
The majority of IRC’s impact may come from their contribution to the evidence-base. In the Ott (forthcoming) rapid evidence review of intervention impact evaluations in forced displacement settings, a significant fraction were produced by IRC. The IRC has completed or is in the process of completing 101 research studies across 32 crisis-affected contexts (including 28 ongoing studies across 23 countries).133 Out of roughly 120 high-quality impact evaluations ever completed in conflict-affected or refugee contexts, IRC have contributed to, or directly conducted, 20 – and are in the process of running 17 more.134
Modelling the value of new information in this space is difficult for two reasons:135
1. It is unclear how cost-effectiveness is distributed in the humanitarian space or among organisations doing direct work in the developing world more generally, and unclear what
130 https://www.rescue.org/page/how-we-use-evidence 131 International Rescue Committee, “Seven Steps to Scaling Cash Relief: Driving Outcomes and Efficiency,” 5. 132 International Rescue Committee, 5. 133 International Rescue Committee, written correspondence 29th May 2018. 134 International Rescue Committee, written correspondence 29th May 2018. 135 For an overview of how to model value of information see Edward C. F. Wilson, “A Practical Guide to Value of Information Analysis,” PharmacoEconomics 33, no. 2 (February 1, 2015): 105–21, https://doi.org/10.1007/s40273-014- 0219-x.
44 — Founders Pledge Forced Displacement
our prior over these should be.136 We don’t think impact is normally distributed and our best guess is that it is distributed log-normally, though we are very unsure about this.
2. It is unclear how responsive the humanitarian sector is to new information. There are some cases, such as the shift to cash transfers, in which the sector has been responsive, though it is difficult to generalise from this case to all other humanitarian interventions.
With these limitations in view, we developed some models trying to estimate the expected benefits of new research on humanitarian interventions. We have low confidence in these models and hope to improve on modelling of new information in the coming year. The models estimated that IRC’s research potentially has significant positive value because it stands to influence some part of the vast amount of money spent on humanitarian aid (now more than $20bn). However, on neither model were the benefits great enough to make IRC’s combined direct work comparable in terms of impact to GiveDirectly. The other direct work would include cash transfers as well as other humanitarian support, which we would expect to be much less cost-effective than cash, which in turn drags down their cost-effectiveness.
Moreover, IRC’s budget in 2017 was large, at around $717m.137 We are therefore not convinced IRC is funding constrained, though IRC told us that demand for their services continually outstrips funding and that their Lebanon and Jordan campaigns are currently underfunded.138
International Refugee Assistance Project
The International Refugee Assistance Project (IRAP) is a US-based NGO established in 2009 that provides legal services to refugees and advocates for policy change to benefit refugees. As discussed, we decided not to properly evaluate their policy work in this report. Through its direct work, IRAP provides pro bono legal services to forcibly displaced people across the world. Its representation relies on a cadre of volunteers—1,200 students from 30 law schools in the United
136 For discussion of the distribution of impact in the charitable space, see Jeff Kaufman, “Effectiveness: Gaussian?,” August 2015, https://www.jefftk.com/p/effectiveness-gaussian; Brian Tomasik, “Why Charities Usually Don’t Differ Astronomically in Expected Cost-Effectiveness,” Essays on Reducing Suffering, September 2017, http://reducing- suffering.org/why-charities-dont-differ-astronomically-in-cost-effectiveness/. See also the comments to Holden Karnofsky, “Maximizing Cost-Effectiveness via Critical Inquiry,” The GiveWell Blog, November 10, 2011, https://blog.givewell.org/2011/11/10/maximizing-cost-effectiveness-via-critical-inquiry/. 137 International Rescue Committee, “Annual Report,” 2017, 37, https://www.rescue.org/resource/international-rescue- committee-annual-report-2017. 138 IRC, written correspondence, May 25th 2018.
45 — Founders Pledge Forced Displacement
States and Canada and pro bono attorneys from over 100 international law firms and multinational corporations—assisting refugees on urgent registration, protection, and resettlement cases.139
IRAP claims to have resettled over 3,800 refugees and displaced persons to 18 different countries, and provided legal assistance through its online advice hotline to more than 20,000 individuals.140 We were generally positive about IRAP’s work, but we are not convinced that they are significantly funding constrained at this time. Over their past five years, their budget has grown at an average of 60%, perhaps driven by the reaction to the refugee policies of the Trump administration.141 Their budget has grown from $1.1m in 2015 to $4.3m in 2018. Their required budget for the financial year 2018 has been fully funded, and a total of $3.2m has been pledged for their 2019 budget of $4.8m.142 They already have a wide range of large philanthropic supporters, so we do not expect them to be significantly funding constrained for the next financial year.
Sanku Project Healthy Children
Sanku Project Healthy Children is a charity carrying out micronutrient fortification in Africa. Until a few years ago, Project Healthy Children only advocated for large-scale micronutrient fortification of staple foods, but the majority of people living in rural and remote areas do not have access to centrally processed food, instead depending on small-scale rural flour mills. Consequently, in 2013 Project Healthy Children rolled out Sanku, an initiative which provides specially developed fortification machines to small-scale flour mills. Since 2015, Sanku has partnered with the UN World Food Program to fortify flour for the Kakuma refugee camp in Kenya.143 Sanku is in the process of expanding this model to reach Burundian refugees in Rwanda, starting with the Mahama refugee camp. We evaluated this project in some depth.
Sanku planned to fortify flour with zinc, iron, folate, and vitamin B12.144 The fortified food would have near 100% coverage within the camps because all refugees in camps consume the flour provided by the mills every day. Sanku’s target is to reach 4m refugees by 2022,145 and the project is funding constrained. 146
139 IRAP, written correspondence, May 16th 2018. 140 IRAP, written correspondence, May 16th 2018. 141 Sarah Vaughan Troyer, written correspondence, May 16th 2018. 142 Sarah Vaughan Troyer, written correspondence, May 16th 2018. 143 Sanku, refugee project concept note, 2018. 144 Conversation with Felix Brooks-Church of Sanku, 26th April 2018. 145 Sanku, refugee project concept note, 2018. 146 Conversation with Felix Brooks-Church of Sanku, 26th April 2018.
46 — Founders Pledge Forced Displacement
There are some studies showing that micronutrient fortification is a highly cost-effective way to improve health and to increase future earnings,147 though we are unsure the accuracy of some of these studies. The reason we decided not to recommend Sanku was that we are concerned about the risk that iron fortification could increase morbidity and mortality from malaria. There is evidence that in malaria endemic areas, unless there is strong malaria prevention and treatment in place including wide distribution of bednets and comprehensive provision of malaria prophylaxis, iron supplementation increases malarial morbidity and mortality,148 potentially outweighing the benefits of reduced anaemia. The risk seems to be mainly driven by providing further iron to already iron replete people.149 There are a number of proposed mechanisms for the link between iron and malaria,150 but all posit that the increased provision of dietary iron is a risk factor for malaria.
Our reading of the evidence is as follows. Insofar as iron fortification also provides additional dietary iron, we should expect iron fortification to increase malaria risk. There are no trials testing directly whether iron fortification increases malaria in the absence of malaria treatment, but this is because it would now be impossible to gain ethical approval for such a trial.151 Note that Sanku disagrees with this reading of the evidence and believes that iron fortification is safe in malaria endemic areas.152
The evidence suggests that malaria risk over the next few years is high enough to make this risk a concern. Throughout the past few years, malaria has been endemic in the Mahama camp and in the surrounding area in Rwanda. The UNHCR writes that:
147 See for example Susan Horton, Venkatesh Mannar, and Annie Wesley, “Micronutrient Fortification (Iron and Salt Iodization)” (Copenhagen Consensus, 2008), http://www.copenhagenconsensus.com/publication/micronutrient- fortification; Rob Baltussen, Cécile Knai, and Mona Sharan, “Iron Fortification and Iron Supplementation Are Cost- Effective Interventions to Reduce Iron Deficiency in Four Subregions of the World,” The Journal of Nutrition 134, no. 10 (October 1, 2004): 2678–84, https://doi.org/10.1093/jn/134.10.2678. 148 Ami Neuberger et al., “Oral Iron Supplements for Children in Malaria-endemic Areas,” in The Cochrane Library (John Wiley & Sons, Ltd, 2016), https://doi.org/10.1002/14651858.CD006589.pub4. 149 Gary M. Brittenham, “Safety of Iron Fortification and Supplementation in Malaria-Endemic Areas,” Nestle Nutrition Institute Workshop Series 70 (2012): 117–27, https://doi.org/10.1159/000337674. 150 Brittenham, 5–7. 151 Conversation with Ami Neuberger, 23rd May 2018. 152 Sanku, document on malaria and iron fortification, shared May 16th 2018.
47 — Founders Pledge Forced Displacement
“Due to Mahama’s location in an insect-infested and swampy area, there are inherent hazards particularly for young children and pregnant women, with a high possibility of contracting malaria and other endemic tropical diseases.”153
We have been unable to find data on trends in malaria prevalence over the last few years in Rwandan refugee camps.154 In 2017, malaria was the leading cause of morbidity in the Mahama camp.155 Sanku have told us that according to UNHCR, in 2017 rates of malaria were as high as 30% in the Mahama camp, though this has fallen to 2% as of 2018 due to indoor residual spraying.156 However, we are not confident that low rates of malaria will persist in camps over the next few years. Malaria was recognised as a serious problem at least in summer 2015,157 but it took two and a half years for effective action to be taken. Managers of refugee camps are resource constrained, and we would not be surprised if malaria increased over the next few years due to changes in funding. For these reasons, our concerns about heightened malaria risk were sufficient to not recommend Sanku Project Healthy Children, though we again note that they disagree with our interpretation of the evidence on the link between iron fortification and malaria.
153 UNHCR, “Burundi Regional Refugee Response Plan,” 2018, 40, https://reliefweb.int/sites/reliefweb.int/files/resources/61371.pdf. 154 A compendium of situation reports on Rwandan refugee camps is available here - https://www.unicef.org/appeals/other_emergencies.html#rwanda 155 UNICEF, “Rwanda Humanitarian Situation Report: Burundi Refugees,” December 2017, 5, https://reliefweb.int/sites/reliefweb.int/files/resources/UNICEF%20Rwanda%20Humanitarian%20Situation%20Report%2 0-%20Jan%20to%20Dec%202017.pdf. 156 Felix Brooks-Church, email correspondence 12th July 2018. 157 UNICEF, “Rwanda Humanitarian Situation Report,” June 2015, https://www.unicef.org/appeals/files/UNICEF_Rwanda_Burundi_Refugees_SitRep_03Jun2015.pdf.
48 — Founders Pledge Forced Displacement